Epidemiology data indicate that regular use of non-steroidal anti-inflammatory drugs (NSAIDs) reduces the incidence of colorectal cancer by approximately 50% in humans. Rodent models of colon cancer also show reductions in tumor formation following NSAID treatment. NSAIDs are believed to act by inhibition of cyclooxygenase (COX-1 and COX-2). Our previous studies have shown a reduction in precancerous lesions in COX deficient Min mice. In the present study azoxymethane is being used to induce colon cancer in wild type and COX deficient mice. Azoxymethane initially induces precancerous aberrant crypt foci followed by adenocarcinomas. Azoxymethane treatment of mice genetically deficient in either COX-1 or COX-2, heterozygous for COX-1 or COX-2, or heterozygous for both isoforms, provides a method of studying COX dependent development of premalignant and malignant lesions. Our studies with the COX deficient mice should provide valuable information about the advantages of selectively inhibiting COX-1 or COX-2 vs the use of dual COX inhibitors.